EPA-650/4-75-024-g
Environmental Monitoring Series
GUIDELINES
FOR QUALITY ASSURANCE PROGRAMS
FOR MOBILE SOURCE EMISSIONS
MEASUREMENT SYSTEMS:
PHASE IV, HEAVY-DUTY GASOLINE ENGINES -
QUALITY ASSURANCE GUIDELINES
U.S. Environmental Protection Agency
Office of Research and Development
Washington, 0. C. 20460
-------
EPA-650/4-75-024-g
GUIDELINES
FOR QUALITY ASSURANCE PROGRAMS
FOR MOBILE SOURCE EMISSIONS
MEASUREMENT SYSTEMS:
PHASE IV, HEAVY-DUTY GASOLINE ENGINES -
QUALITY ASSURANCE GUIDELINES
by
Harold Wimette, Rod Pilkington, and Tom Kelly
Olson Laboratories, Inc.
421 East Cerritos Avenue
Anaheim, California 92805
Contract No. 68-02-1740
ROAP No. 26BGC
Program Element No. 1HA327
EPA Project Officers:
R. C. Rhodes
Quality Assurance and Environmental Monitoring Laboratory
Research Triangle Park , North Carolina 27711
and
C. Don Paulsell
Office of Program Management
Ann Arbor, Michigan 48105
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Mobile Source Air Pollution Control
and
Office of Research and Development
Washington, D. C. 20460
June 1975
-------
EPA REVIEW NOTICE
This volume has been prepared by Olson Laboratories, Incorporated
consistent with the Environmental Protection Agency Quality Assurance
principles and concepts and with the Environmental Protection Agency Mobile
Source Testing Practices at Ann Arbor, Michigan.
The guidelines and procedures are generally applicable to mobile
source testing operations and are intended for use by those engaged in such
measurement programs
It is requested that recipients and users of this document submit any
comments and suggestions to the Project Officers.
Mention of trade names or commercial products does not constitute
Environmental Protection Agency endorsement or recommendation for use.
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environ-
mental Protection Agency, have been grouped into series. These broad
categories were established to facilitate further development and applica-
tion of environmental technology. Elimination of traditional grouping was
consciously planned to foster technology transfer and maximum interface
in related fields. These series are:
i. ENVIRONMENTAL HEALTH EFFECTS RESEARCH
2 . ENVIRONMENTAL PROTECTION TECHNOLOGY
3. ECOLOGICAL RESEARCH
4. ENVIRONMENTAL MONITORING
5. SOCIOECONOMIC ENVIRONMENTAL STUDIES
6. SCIENTIFIC AND TECHNICAL ASSESSMENT REPORTS
9. MISCELLANEOUS
This report has been assigned to the ENVIRONMENTAL MONITORING
series. This series describes research conducted to develop new or
improved methods and instrumentation for the identification and quanti-
fication of environmental pollutants at the lowest conceivably significant
concentrations. It also includes studies to determine the ambient concen-
t rations of pollutants in the environment and/or the variance of pollutants
as a (unction of time or meteorological factors.
This document is available to the public for sale through the National
Technical Information Service, Springfield, Virginia 22161.
Publication No. EPA-650/4-75-024-g
-------
FOREWORD
All mobile source testing facilities have some elements
(activities) of a quality assurance system built into their routine
testing operations. These activities may not have been identified
and/or integrated into a formal quality assurance program. It is the
objective of these guidelines to provide guidance to both (1) facilities
which desire to organize an integrated quality assurance program; and
(2) facilities which may have already organized towards an integrated
quality assurance program, but may desire to review their program as a
result of the recommendations and suggestions included in these guide-
lines. The extent of implementation of these guidelines will depend
upon the requirements of each individual test facility.
111
-------
EXECUTIVE SUMMARY
Many of the principles of quality assurance described in
Phase II, which detailed guidelines for the measurement of heavy duty
diesel engine emissions, are generally applicable to all measurement
systems. As a supplement to the Phase II report, this document examines
the technological aspects of heavy duty gasoline engines, with reference
to applicable functions and procedures.
The measurement system used for heavy duty gasoline engines
is described in detail in Volume I and test procedures used by the EPA,
Ann Arbor facility to meet applicable requirements of the Federal
Register for the 1975 model-year, appear in Volume II.
v
-------
TABLE OF CONTENTS
FOREWORD iii
EXECUTIVE SUMMARY v
Section
1 INTRODUCTION 1-1
1.1 Objective and Scope of Guidelines 1-2
1.2 Formation of Quality Assurance Guidelines 1-2
1.2.1 Section 1 Introduction 1-2
1.2.2 Section 2 Organizing for Quality 1-3
1.2.3 Section 3 Measurement System-Analysis 1-3
1.2.4 Section 4 Guidelines for Performance,
Audits and Maintenance Procedures 1-3
1.2.5 Section 5 Quality Assurance Guidelines for
Documentation of the Measurement System 1-3
1.2.6 Section 6 Application of Statistical Quality
Assurance Methods to the Emission
Test System 1-3
1.2.7 Section 7 Analysis of Variability in the
Measurement of Emissions from Light Duty
Diesel Engines 1-3
1.2.8 Section 8 Quality Assurance System
(On Site) Survey 1-3
1.2.9 Appendices 1-3
2 ORGANIZING FOR QUALITY 2-1
2.1 Operations Management 2-1
2.1.1 Quality Assurance Management 2-1
2.1.2 Emission Test Facility Management 2-1
3 MEASUREMENT SYSTEM ANALYSIS 3-1
3.1 Applicable Federal Register Procedures 3-1
3.2 Elements of a Measurement System for Heavy Duty
Diesel Engine Emission Measurement 3-2
4 GUIDELINES FOR PERFORMANCE AUDITS AND MAINTENANCE
PROCEDURES 4-1
5 QUALITY ASSURANCE GUIDELINES FOR DOCUMENTATION OF THE
MEASUREMENT SYSTEM 5-1
Vil
-------
6 APPLICATION OF STATISTICAL QUALITY ASSURANCE METHODS
TO THE EMISSION TEST SYSTEM 6-1
7 ANALYSIS OF VARIABILITY IN THE MEASUREMENT OF
EMISSIONS FROM LIGHT DUTY DIESEL ENGINES 7-1
7.1 Variables Associated with the Measurement System-Test
Cell Equipment and Instruments 7-2
7.1.1 Dynamometer Operation 7-3
7.1.2 Relative Humidity 7-3
7.1.3 Barometric Pressure 7-4
7.1.4 Fuel Measurement and Control 7-4
7.2 Variables Associated with the Measurement of Gaseous
Emissions 7-6
7.2.1 Variables Associated with the Analytical System . . 7-6
7.2.2 Variation Assciated with the Data Reduction .... 7-7
7.2.3 Computer 7-8
7.3 The Engine as a Source of Variability 7-9
7.4 Measurement of Variability in the Emission
Measurement System 7-9
7.5 Quality Assurance and Test Variability 7-13
8 QUALITY ASSURANCE SYSTEM SURVEY 8-1
9 REFERENCES 9-1
Appendices
A, B,
and C Refer to Phase II Report
Vlll
-------
LIST OF FIGURES
Figure No. Page
7-1 Effect of Ambient Temperature on Exhause Emissions
During the CVS Cold-Start Test ,.,,., 7-^10
LIST OF TABLES
Table No, Page
3-1 Supart H-Emission Regulations for New Gasoline-Fueled
Heavy Duty Engines 3-2
3-2 Heavy Duty Gasoline Engine Emission Measurement
System 3-4
4-1 Federal Register Specifications Heavy Duty Gasoline
Engine Emission Measurements - Subpart H 4-2
7-1 Effect of Barometric Pressure and Humidity on Exhaust
Emissions . ,.,..,.,,. 7-11
7-2 Summary of Test Variables and Methods Used For
Their Control , 7-14
IX
-------
Section: 1(HD)
Revision: 0
Date: June 1975
Page 1 of 4
Section 1
INTRODUCTION
The Quality Assurance Staff of the EPA Quality Assurance and
Environment Monitoring Laboratory, Research Triangle Park, North Carolina,
is responsible for the administration of a Quality Assurance Program
for air measurement systems resulting from the implementation of the
Clean Air Act. Standards for the emissions for light and heavy duty
mobile sources have been promulgated and procedures published for the
measurement of their emissions and certification. Quality Assurance
guidelines, however, have not been previously specified of these mobile
source emission measurement procedures. Such quality assurance programs
are necessary to assure the integrity of the data resulting from these
tests. This report presents guidelines for quality assurance programs
for measurement systems used in mobile source testing according to the
applicable requirements of the Federal Register for the 1975 model-
year.
The guidelines for the Quality Assurance Program for mobile
source measurement systems were prepared in four phases.
o Phase I - For light duty gasoline-powered vehicles (cars
and trucks)
o Phase II - For heavy duty diesel engines
o Phase III - For light duty diesel-powered vehicles (cars
and trucks)
o Phase IV - For heavy duty gasoline engines
This document presents the guidelines for implementing a
Quality Assurance Program for the measurement of emission from heavy
duty gasoline engines (Phase IV). Guidelines for the other phases were
reported in separate documents.
-------
Section: 1(HD)
Revision: 0
Date: June 1975
Page 2 of 4
1.1 OBJECTIVE AND SCOPE OF GUIDELINES
These guidelines provide information on general quality
methods which may be used in emission testing. They were primarily
designed for use by management and supervisory personnel involved in
the development or operation of quality programs. Upper managment may
use the guidelines to evaluate the quality programs which presently
exist within their own laboratory or organization.
The measurement system for heavy duty gasoline engines consists
of the testing, calibration and analytical requirements, the operational
and measurement data obtained. The primary objective of this program
was to analyze this system and apply the principles and techniques of
modern quality assurance systems to the total testing process to assure
the validity and reliability of the tests and the resulting test data.
Many of the guidelines and test procedures described in
phase II of this program are directly applicable to the heavy duty
gasoline engine emission measurement system. Consequently, the objective
of this supplement is to provide additional information and procedures
required specifically for heavy duty gasoline emission tests.
1.2 FORMATION OF QUALITY ASSURANCE GUIDELINES
In order to identify those areas requiring special definition
for Phase IV, the report for Phase II was reviewed along with the
available information concerning heavy duty gasoline engine emission
test procedures to determine necessary revisions or modifications of
the Phase II documents. Sections and paragraphs requiring revision are
numbered identically to the original document to facilitate cross
reference. Sections applicable in their entirety are noted as such.
The quality assurance guidelines for heavy duty engine emission
measurement systems are contained in Sections 1 through 8, with all
references appearing in Section 9. A summary of the contents of each
section is as follows:
1.2.1 Section 1 - Introduction
A description of the background, objective and organization
of the guidelines,
1.2.2 Section 2 - Organizing For Quality
A typical Quality Assurance Organization is presented.
Quality functions are identified and the various key elements of a
quality program are described.
-------
Section: 1(HD)
Revision: 0
Date: June 1975
Page 3 of 4
1.2.3 Section 3 - Measurement System Analysis
A description of the measurement system defining the equipment,
test procedure specifications and tolerances, quality provisions and
other requirements necessary for emission testing of heavy duty gasoline
engines.
1.2.4 Section 4 - Guidelines for Performance Audits and Maintenance
Procedures
General guidelines are presented for performance inspection
and maintenance of instruments and equipment used in the measurement
systems. Preventive maintenance programs are described for increasing
the reliability and efficiency of the test equipment.
1.2.5 Section 5 - Quality Assurance Guidelines for Documentation
Of The Measurement System
Guidelines for the development of a documentation system are
presented with representative forms, a description of the manuals, data
recording, and failure analyses used by a quality assurance program.
1.2.6 Section 6 - Application Of Statistical Quality Assurance
Methods To The Emission Test System
Basic statistical techniques such as control charts, analysis
of variance and data validation as applied to a quality system are
described.
1.2.7 Section 7 - Analysis Of Variability In The Measurement Of
Emissions From Light Duty Vehicles
Sources of variability are identified and, where possible,
quantified to show their effect on the data.
1.2.8 Section 8 - Quality Assurance System (On Site) Survey
A procedure and survey form for conducting a quality assurance
survey of a laboratory performing heavy duty emission testing is presented.
1.2.9 Appendices
Statistical techniques and nomenclatures appear in Appendix
A-l. Appendix A-2 contains control chart multiplication factors.
Appendices B-l and B-2 include a glossary of terms and a list of
-------
Section: 1(HD)
Revision: 0
Date: June 1975
Page 4 of 4
abbreviations commonly used in the measurement system. Appendix C of
Volume I contains general Quality Managment Procedures (QMP) which
define those functions identified as being necessary in a quality
assurance program.
-------
Section: 2(HD)
Revision: 0
Date: June 1975
Page 1 of 1
Section 2
ORGANIZING FOR QUALITY
2.1 Operations Management
2.1.1 Quality Assurance Management
2.1.2 Emission Test Facility Management
Administrative procedures and Quality
Assurance functions are identical with
those of Phase II
-------
Section: 3(HD)
Revision: 0
Date: June 1975
Page 1 of 4
Section 3
MEASUREMENT SYSTEM ANALYSIS
A total Measurement System can be defined as an orderly
arrangement consisting of the analytical method, the test sampling
procedure, the instruments or analyzers, the supporting functions, the
management organization, and the technicians or personnel involved in
performing specific functions within the system. Applying this defini-
tion to the measurement system for heavy duty gasoline engine emissions
the process is composed of:
o The test procedure defined by the Federal regulations
o The preparation of the gasoline engine for the emissions
test
o The exhaust emission sample transfer and analytical
console consisting of NDIR instruments for the measurement
of carbon monoxide (CO) and nitrogen oxides (NO ), and
hydrocarbons (HC)
o The Laboratory Operations management including Test
Operations and Support Operations
This measurement system was subjected to a functional analysis
to determine and define the basic elements which require attention in a
total quality assurance program.
3.1 APPLICABLE FEDERAL REGISTER PROCEDURES
Measurement system for the quality assurance guidelines and
procedures have been developed as defined in the Federal Register. The
Federa.1 Register which defines the measurement systems covered by this
document is Volume 40, Number 40, dated February 27, 1975, pages 8482
to 8495.
The paragraphs which define the test procedure for heavy duty
gasoline engines are summarized Table 3-1.
13
-------
Section: 3(HD)
Revision: 0
Date: June 1975
Page 2 of 4
Table 3-1. SUBPART H-EMISSION REGULATIONS FOR
NEW GASOLINE-FUELED HEAVY DUTY ENGINES
85.701 General applicability
85.702 Definitions
85.703 Abbreviations
85.774-1 Emission standards for 1974 and later model-year
engines
85.774-7 Service accumulation and emission measurements
85.774-9 Test procedures
85.774-10 Gasoline fuel specifications
85.774-11 Dynamometer operation cycle and equipment
85.774-12 Dynamometer procedures
85.774-13 Sampling and analytical system measuring exhaust
emissions
85.774-14 Information to be recorded
85.774-15 Calibration and instrument checks
85.774-16 Dynamometer test run
85.774-17 Chart reading
85.774-18 Calculations
3.2 ELEMENTS OF A MEASUREMENT SYSTEM FOR HEAVY DUTY DIESEL ENGINE
EMISSION MEASUREMENT
A requirement of a total Quality Assurence Program is to
maintain control at all important stages of a process. In this measure-
ment system, an analytical process, it is necessary to first identify
its functional elements. In order to categorize these elements and the
related tasks, the measurement system was divided into the following
operations:
14
-------
Section: 3(HD)
Revision: 0
Date: June 1975
Page 3 of 4
o Engine receiving inspection
o Engine-dyno preparation
o Engine-dyno checkout
o Engine test cycle preconditioning
o Engine test cycle
o Sampling system
o Instrument calibration
o Interference check CO/H O
o Measurement of gaseous emissions
o Data collection and reduction
A summary matrix of the emission measurement procedure for gaseous
emissions is presented in Table 3-2. The overview represented by this
matrix was designed to give a general understanding of the process
involved in exhaust emission testing. However, it was not intended to
include every detail required for this measurement. The information
discussed in this table consist of:
o A brief description of the tasks
o Applicable Federal Register paragraphs
o Applicable EPA, Ann Arbor, test procedure numbers
o Specification and tolerance included in the Federal
Register, and from engineering practices such as the SAE
recommended practices
o Quality provisions
o Invalid test (determination)
o Corrective action required
o Training and skill level required
15
-------
Section:
Revision:
Date:
Page 4 of4
3(HD)
0
CO
>>
co
EH
2
D
CO
2
O
H
co
CO
H
o
§
£
H
g
i
B
.Q
(0
EH
3 UJ
g"
z J
Ss
- o
Is
" 3:
^
£1
_J */•
^1
t_
Is
:<>•
3 j
:s
S°
5
SPECIFICATION OR
TOLERANCt-FED. REG
.a
«^
a.
|
Q.
a
LL.
UJ
CC
X ^
IS
Is
C r-
01 L. 2
l(- C U
O 0 01
cn^o c u
5^51
lift
i
i/i OJ
U I- i- Oi
(T5 3 *J C
SID O C
>«- a. o>
s
Q. 4-1 QJ
!!!
4> a. c1
3 a.
— .CO.
c
0
S
a.
c
c -
>
II
<
L.
° O •
T) L. K
S'ii'T
III*
o ^
So
*O r— «• O
a" 7^0.
u. tsj co *—
c
UJ f- c 19
ro •—
ai h. i—
JT 01 OJ -0
*-> L. cn-^
Engine shipped to
from the manufactu
inspected for dama
conformance to ins
en o
c <-
01 > U
•~ 'S 0.
cT (iJ c
-i
zi
0
*« 01
i? .
Ill
1
Oi
Q
L.
S
,
Incomplete/ Imp
installation
->,
Is
*j a.
i — CO)
«t
cv
•
0
s
*£ <^
Dynamometer capable
maintaining constan
speed -100 RPM from
throttle to closed
throttle motoring
C
7?*?
S XS
IrtlOti
"*""
T) .C O
fcw-w
en E -w
r^|
The engine and pre
stand are installe
test cell and atta
the dynamometer
|
ID
c c
ti
».
•*- >>
o •—
en ri §
•— c
c D.
U 0
ill
o 2
4s
«t un
i1
UJ
fcg
S £
L.
0 S
f 1
-O C
— CT
35
u
5.
i5
S-S
TO U
*t IU
VI »-•
Adjustments on engi
sre 1 united to
unscheduled mainte-
nance listed in
Ref. Para.
CM
5 o
S S
LT1 Q.
co »—
"S -G OI 1 £
*J C L. D (q
fQ O ra C u
The engine is oper
at idle and load c
tions to check for
proper operation.
Engine parameters
checked against ma
facturer's specifi
1**
° o
11
«
"1
o -^
Si «i
"So
i|
i
8
Is
Out of liuit
soak time/
Ternp/ Cycle
§ ^ 3
3 -^
CT TJ >1 C
i. O ^ O
? u "o "?
a
I
^ E
3
•- U
t. r—
II
£
Engine must reach
lormal operating
:onditions, soak tn
1-2 hours temp ,
50-86°F
2
^_"
S S
S 2
M 1—
i t- M
a> o ai 4-> ia
u u c
||
S* **ts
r- O» *-, O « •"-»
*J -o a* e: —
'> sog-S-S m
Sh- t -So
tj o o £ E ^p
c
o
•3^
il
o
CO tl
k.
s
n
is
!/icZ
CD I—
CJ
a,s, S
Iflf
^ O OI
c:a*c o x
< — n u S
^
OI
en u
£3
^
C •
£ t->
£ 2,
i:
<9
O -
Is
K
1
W. 4->
> «
t- OJ
ss
i , fel
o. o x -o
e u o t.
System leak/co
ent failure/ in
rect sample fl
pressure/calib
overdue
•? i o i-
•o u o v- c
j= ° S. ^ 2
** c o
1o i. o *-«^
o S. ° >, c
L. jJ in o
Q. T3 C
< 5 loS
C O
** 0*~
w 2°'i
in u 4)
i^» ^i n >»
C 1- qj o
Z J u.
e ? tn £
^ ^ a ^
>
0*^ Ho
ego.
•a O. CMC
*- O.QO Q.
O o Q.
, * f- M C
Comprised basicall
5 NDIR instruments
Sample pumps, refr
tion, flow control
pressure and flow i
en
= s
Q.*J
»
01
I
-f-
19 t-
^ -li C
u jS U
.
JX X C O 3
U. U- O C L.
"- T3 - U 3 '
•— 0 • 2 Q
>q 0 X 0 O £ C
£
£ss
r*. (M r^
C0»-t-
X
^=0, S£
*J C *J , £ ^ >
Calibration gases
concentrations var
across the instrum
range are used to
the instrument cur
data reduction
c
^s
ei
-
Oi
1
ID
fc.
01
^s~i
5 "°
CT i- OI
I'll
Excessive inte
fere nee from
C02/H20
^
1
=-,
*J
1
il^'ji
t lO O
uw O in . *'
O X I J=
^
rC S
^ncT
»fr
^
£ a*
SfefcS
1_ tt- *J x
JJ ro 0
«^ >> X •"-
The Hydrocarbon in
are checked monthl
interference from
vapor and carbon d
01
I p -*
-^ 5. u
si!
CO
U 1- O 3
OI «- t-
i- 3 11 *->
i_ cr en IA
o 4i "o c
^ I U
QJ 1 S C
> AC O
o c t. -
1C •- i- U *J
« 10 S 01 E
|LS
E l- c OJ
si"! 5
^ >, c o>
<_) i/i •- L.
3 >*-
X C C L 3
c ol
o 'S
1C C
•o o *-•
"m ^o "3
S 3 o-
*j en
c - o
o"°~
C 3
>, l> cn
f— Q. C
I!!5
cS"l
Q. C 8 OJ
*~ 2 *
^
^*
^s
^2
:3t-
*j m
U *- 1 *J
A portion of the e
gases are analyzed
tinuously from the
idle thru four nin
cycles ending with
engine at idle.
.£
££,
o!
^J
I
i.
li
E
||fL
^ IS?I
i.
C ^1 ^ •§
° ~ 4-1
5 £ o" £
^ OJ 3*0
ill!
, = {
^^: "3 ° ^
0 3 01 -0
^^*- CLIj. 0|
en o *-> "^ a.
E^ltl
w>
sea
^^
31-
^ itl I
HJsjjj
!§!„_
2W 01 "- I.
•*j u o JT
§O) TJ Q. vl
E * ^ E
3 Li CT>
u *> o j<"~'
O C t. Ul
J--g^g
•t- *. UV 0 -f-
0 0 3 *-• «1
C
o
ll
o
16
-------
Section: 4(HD)
Revision: 0
Date: June 1975
Page 1 of 5
Section 4
GUIDELINES FOR PERFORMANCE AUDITS
AND MAINTENANCE PROCEDURES
The guidelines presented in this section are applicable to
both measurement systems. A listing of additional specifications for
heavy duty gasoline engine testing are presented in Table 4-1.
19
-------
Section: 4(HD)
Revision: 0
Date: June 1975
Page 2 of 5
Table 4-1. FEDERAL REGISTER SPECIFICATIONS HEAVY DUTY GASOLINE
ENGINE EMISSION MEASUREMENTS - SUBPART H
REFERENCE
PARAGRAPH
85.702
85.774-1
S5. 774-7
85.774-10
85.774-11
PROCEDURE OR EQUIPMENT DESCRIPTION
Zero Hour Engine-Defined
Heavy Duty Vehicle
Heavy Duty Engine
Emission Standards 1974 and later
model-year engines
Result of Emission Tests
Emission and Durability 125 hour
test point
Gasoline Specifications
Dynamometer Specifications
Exhaust System
Cooling
SPECIFICATION OR TOLERANCE
That point after normal assembly line operations and adjustment
and before one additional operating hour has been accumulated.
Rated at more than 6,000 Ibs. GVW or designed for transportation
of property or more than 12 persons .
One used for motive power is a heavy duty vehicle.
1. Hydrocarbons plus oxides of nitrogen
per brake horsepower hour.
2. Carbon monoxide. 40 grams per brake
3. No crank case emissions are allowed.
(as NO ) - 16 grams
horsepower hour.
Reported using two places to the right of the decimal point,
rounded off according to ASTM E29-67.
±8 hours of 125 hour multiple.
ASTM
ITEM DESIGNATION LEADED UNLOADED
Octane, Research minimum D1656. . . .
Pb. (organic), grams/U.S. gals
Distillation range D86
IBP . °F
10 percent, °F
50 percent, F
90 percent, F
EP , F (maximum)
Phosphorus grams/U . S . gals max
RVP, pound D323
Hydrocarbon composition. D1319 . .
. ..100 96.
...1.4 min 0 . 00-0 . 05 .
. . .75-95 75-95.
. ..126-135 126-135.
. . .200-230 200-230.
. ..300-325 300-325.
. . .415 415.
. ..0.10 0.10.
. ..0.01 0.005.
. . .8.0-9.2 8.0-9.2.
. . .10 10.
. . .35 35.
Capable of maintaining speed ±100 RPM from full throttle to
closed throttle
Chassis - type or equivalent
Radiator to maintain typical engine operating temperatures .
Fixed speed fan may be used.
20
-------
Table 4-1. FEDERAL REGISTER SPECIFICATIONS HEAVY DUTY GASOLINE
ENGINE EMISSION MEASUREMENTS - SUBPART H (Continued)
Section: 4(HD)
Revision: 0
Date: June 1975
Page 3 of 5
REFERENCE
PARAGRAPH
PROCEDURE OR EQUIPMENT DESCRIPTION
SPECIFICATION OR TOLERANCE
85.774-11 Test Cycle
MANIFOLD TIME IN CUMULATIVE WEIGHTING
SEQUENCE NO. MODE VACUUM MODE-SECS. TIME-SECS. FACTORS
1 Idle 70 70 0.232
2 Cruise. 16" Hg. . 23 93 .077
3 PTA . . 10" Hg. . 44 137 .147
4 Cruise. 16" Hg. . 23 160 .077
5 PTD . . 19" Hg. . 17 177 .057
6 Cruise. 16" Hg. . 23 200 .077
7 FL. . . 3" Hg. . 34 234 .113
8 Cruise. 16" Hg. . 23 257 .077
9 CT. . . . .... 43 300 .143
Dynamometer
Idle Speed
Closed Throttle Mode (CT)
Part Throttle Mode (PTD)
Full Load Mode (FL)
Operated at a constant speed of 2000 RPM ±100 PRM, ±200 RPM is
allowed during first 4 seconds of each mode.
Idle speed at manufacturer's recommended engine speed.
2000 RPM ±100 RPM
Specified manifold vacuum or closed throttle if vacuum cannot
be obtained.
Specified manifold vacuum or wide open throttle if vacuum
cannot be obtained
85.774-13 Analytical Equipment
85.774-15
Consists of the following basic components sample, probe,
refrigerated bath, filters, pumps, pressure gauges, water
traps, drier for NO, flowmeters and appropriate values and
fittings.
Instruments:
CO - NDIR
Range 0-10 percent
Range 0-16 percent
Range 0 - 4000 PPM
0-1000 PPM hexane equivalent
0-10,000 PPM hexane equivalent
Lower operating ranges may be used as required.
CO- - NDIR
NO - NDIR
HC - NDIR(2)
85.774-14 Recorders-Analyzers, manifold
vacuum and Engine RPM
Automatic 1 second interval marker or preprinted chart paper.
Correct chart speed must be verified and the charts for each
run.
85.774-15 Instrument Calibration
Zero Gas
Flow Rates
Calibration Gases
Every 30 days
Air or nitrogen-impurity concentrations should not exceed
10 PPM NO.
10 C.F.H. for HC and NO
5 C.F.H. for CO and CO
Accuracy ±2 percent of nominal value. Diluted prepurified N .
21
-------
Section: 4(HD)
Revision: 0
Date: June 1975
Page 4 of 5
Table 4-1. FEDERAL REGISTER SPECIFICATIONS HEAVY DUTY GASOLINE
ENGINE EMISSION MEASUREMENTS - SUBPART H (Continued)
REFERENCE
PARAGRAPH PROCEDURE OR EQUIPMENT DESCRIPTION
85.774-15 Calibration Gases-Required Cone
Concentrations
SPECIFICATION OR TOLERANCE
entration determined with ±2 percent of true value.
CO and CO
Low range High range NO ana- analyzers-Blend
HC analyz- HC analyz- lyzer- of CO and CO
er-Hexane . er-Hexane NO containing
equivalent equivalent CO plus CO
Mode Mode
ppm ppm ppm percent percent
100 600 250 0.5 16.0
200 1,000 500 1.0 15.0
300 1,500 750 2.0 14.0
400 2,500 1,000 3.0 13.0
600 4,000 1,500 4.0 12.0
800 6,000 2,000 6.0 10.0
1,000 8,000 2,500 8.0 8.0
10,000 3,000 10.0 6.0
3,500
4,000
The hexane equivalent of propane when used as the
normalizing gas for calibrating nondispersive infrared
analyzers is prescribed to be 0.52 (propane concen-
tration X 0.52 = hexane equivalent concentration).
Minimum storage temperature of the cylinders shall be
60 F; minimum use temperature shall be 68 F.
85.774-15 Hydrocarbon-NDIR Response to C0_ 100 percent CO response less than 0.5 percent if greater
recharge filter cell. If this does not correct problem,
replace detector.
Response to HO Saturated nitrogen at ambient temperature. Low range - if it
exceeds 5 percent of full scale (75 F) , replace detector.
High range not to exceed 0.5 percent of full scale.
85.774-15 Instrument Check Daily - 2 hour warm up - span before and after test - span
after test must repeat within ±2 percent of full scale.
Span gas - 80-100 percent accuracy ±2 percent of true value
gain - shift in excess of ±3 percent requires analyzer to be
retuned and analyzer curve checked. Record actual concen-
trations on chart.
22
-------
Table 4-1. FEDERAL REGISTER SPECIFICATIONS HEAVY DUTY GASOLINE
ENGINE EMISSION MEASUREMENTS - SUBPART H (Continued)
Section: 4(HD)
Revision: 0
Date: June
Page 5 of 5
REFERENCE
PARAGRAPH
PROCEDURE OR EQUIPMENT DESCRIPTION
SPECIFICATION OR TOLERANCE
85.774-16 Dynamometer Test Run
Preconditioning: Service accumlation cycle or 9 mode cycle
run until normal operating conditions
reached.
The engine shall not be exposed to precipitation or
condensation after preconditioning.
Soak: Min. of 1 hour, Max. 2 hours at 60°-86°F.
Sampleline; 2 feet into the tail pipe, dual exhaust
requires two probes with no more than 4 inches
variation in length. Same material, insertion,
diameter and configuration.
Chart Speed: 6 inches per minute.
Hydrocarbon: At the end of the test the HC concentration
Hang Up shall drop to 5 percent or less of full scale
within 10 sec., and 3 percent or less of full
scale within 3 minutes while being purged
with zero gas.
85.774-17 Dynamometer Test Run
Time: ± 2 seconds for CT mode.
Manifold; ±0.3 in. Hg-cruise and PTD mode
±0.2 in. Hg-PTA and FL mode during the 10 seconds
of the mode.
Speed: ±200 RPM, first 4 seconds each mode, ±100 RPM for
remainder.
85.774-17 Chart Reading - Manual
All analyzer traces for 3, 10, 16, 19 in Hg and idle modes.
Divide last 2 seconds into min. 3 segments and read to within
0.5 percent of full scale - average readings.
Initial idle mode used for warmup and cycles 1 and 2, final
idle mode used for cycles 3 and 4.
Closed throttle mode - all traces - divide into min. 43
segments of equal length and determine deflection within
0.5 percent of full scale - average concentrations values.
85.774-17 Chart Reading Computer
(f) Direct computer analysis of analyzer output may be utilized
provided that the analysis is sufficiently similar to the above
procedures to result in comparable data results and the analyzer
output is continuously recorded at a chart speed of at least
3 inches per minute with an automatic marker being used to
identify the time intervals during which data are accepted by
the computer for processing.
85.774-17 Brake Horsepower
Horsepower for the idle and closed throttle mode shall be
defined as zero for calculation purposes, negative vales
are not used.
23
-------
Sections: 5(HD)
6(HD)
Revision: 0
Date: June 1975
Page 1 of 1
Section 5
QUALITY ASSURANCE GUIDELINES
FOR DOCUMENTATION OF THE MEASUREMENT SYSTEM
Section 6
APPLICATIONS OF STATISTICAL
QUALITY ASSURANCE METHODS TO THE EMISSION TEST SYSTEM
The above two sections are directly applicable to both measure-
ment systems. The general guidelines described for documentation and
statistical methods may be incorporated into a quality system for all
mobile source emission testing. When establishing a quality plan for a
particular mobile source testing facility, these two sections of the
Phase II report should be consulted for guidance in such areas as
control of procedural manuals, recording of results, processing and
audit control emission data, initiating control charts, and the implemen-
tation of corrective action procedures.
29
-------
Section: 7(HD)
Revision: 0
Date: June 1975
Page I of 14
Section 7
ANALYSIS OF VARIABILITY IN THE MEASUREMENT
EMISSIONS FROM HEAVY DUTY GASOLINE ENGINES
A knowledge of specific variables significantly affecting the
data is a prerequisite for achieving a predetermined goal, improving
data reliability and detecting bias factors in the system. These
variables are either determinate or indeterminate. Determinate variables
may be objectively studied by engineering evaluation of the test procedure
and statistical analyses of the data. The nature of indeterminate
variables requires them to be evaluated subjectively. Indeterminate
variables are usually estimated through experience with the measurement
system.
The measurement system for heavy duty gasoline engines consists
primarily of an engine-dynamometer test cell and an NDIR analytical
console for the measurement of carbon monoxide (CO), carbon dioxide
(CO ), hydrocarbons (HC) and nitrogen oxides (NO ).
The test sequence involves the measurement of HC, CO, CO and
NO during several steady-state modes run on an engine dynamometer
which are designed to simulate a truck driving pattern in a metropolitan
area. Tolerances and specifications for instruments, equipment and
test cycle appear in Table 3-2 and Table 4-1 of this report.
In addition to engine inconsistencies, certain measurement
system variables have been established as prime source of error.
Efforts to reduce these variables include the use of instruments and
calibration standards having increased precision and accuracy, and
improvement in methods of sampling gaseous emissions.
This section of the guidelines discusses the methods used to
identify these major sources of variation, to quantify the effect of
the determinate variables, and to define the role of quality assurance
in the effort to reduce test variation. However, very little information
has been published concerning the test error associated with the heavy
duty gasoline engine measurement system.
Presently there are some major differences between the light
duty and heavy duty analytical systems. The heavy duty gasoline analytic
cal system requires NDIR measurements for hydrocarbons and nitric
oxide. In light duty emissions these testing instruments have been
replaced by the Flame lonization Detector for hydrocarbons and the
Chemiluminescent -
-------
Section: 7(HD)
Revision: 0
Date: June 1975
Page 2 of 14
with the latter instrument to convert NO2 in the sample to NO. It is
likely that the CL method will be specified in the near future in heavy
duty testing due to undesirable characteristics of the NDIR system,
such as interference from other gases. Until such time as these changes
become permissible by Federal Regulations, the manufacturer must continue
to certify their engines by the present procedure.
7.1 VARIABLES ASSOCIATED WITH THE MEASUREMENT SYSTEM-TEST CELL
EQUIPMENT AND INSTRUMENTS
The primary sources of test-error in the measurement system
are the:
o Dynamometer
o Humidity Measurement
o Barometer
o Ambient Conditions
o Fuel Measurement
o Analyzer
o Sampling System
o Calibration of Instruments
o Zero Gas Purity
o Working Standard
o Operator
o Computer
o Engine
Engine, dynamometer, ambient condition and operator variability
are interrelated variables, all of which usually vary if any one of
them is changed. For example, changes in ambient conditions can affect
operation of engine control systems.
Most of the other variables are determinate, but little
information is available as to their actual contribution to test error.
The gasoline engine measurement system is presently under study at the
EPA, Ann Arbor facility.
The light duty measurement systems has been studied extensively
by the EPA and the major automobile manufacturers, (References 7-2,
7-3, 7-4, 7-5, 7-6, 7-7).
Some of the variables listed are comparable to those associated
with light duty vehicle testing (Reference 7-1). For example, calibra-
tions of the analyzers are performed using the same procedure and
checks. The estimate for the light duty analytical system calibration
coefficient of variation is 1 percent and would be expected to be
nearly the same for heavy duty gasoline emission measurements.
34
-------
Section: 7(HD)
Revision: 0
Date: June 1975
Page 3 of 14
7.1.1 Dynamometer Operations
The dynamometer speed and torque meters are calibrated at
monthly intervals. The associated recorders are checked and aligned
with the meters prior to each test. The RPM output is checked with a
primary standard such as a strobotac with a standard accuracy of
±1 percent. The torque meter is usually calibrated using weights
placed on the torque beam. The weights should be traceable to the NBS
and should have an accuracy of ±0.1 percent. The test variability
associated with the operation of the dynamometer would be difficult to
assess without an extensive program involving several operators, dyna-
mometers and engines.
Computer-controlled dynamometers have become quite popular.
However, thes.e systems require manual intervention, with its associated
inherent variability, to program the computer each time an engine is
installed in the test cell.
The test cycle is operated according to engine RPM and manifold
vacuum within the specified tolerance for a particular mode.
Speed/torque/manifold vacuum traces are made for each test and those
not meeting the tolerance are voided and the test is rerun.
An assessment of this dynamometer variable could be achieved
by performing ten consecutive tests (cycles) on a well-preconditioned
engine using the same computer program or operator followed by ten
consecutive tests during which the RPM and manifold vacuum varied
within the specified limits during the cycle, using different operators
if possible.
This data would give an objective assessment of the variability
due to dynamometer operation. Naturally, this would be valid only for
that particular engine-dynamometer combination. The test would be
repeated with different engine-dynamometer combinations each time a new
engine is installed. A coefficient of variation control chart as
described in Section 6.2.5.3 could be established using the collected
data. The analysis of variance technique, also discussed in Section 6,
could be applied to determine significant differences between the
various combinations.
7.1.2 Relative Humidity
Test error in the determination of relative humidity could be
significant when used to calculate the correction factor for NO . The
correction factor is applied to adjust for the differences in NO
emissions as ambient relative humidity changes. This correction factor
reduces the variability of the NO emission data by normalizing to a
standard relative humidity of 75 grains of moisture per pound of dry
air.
35
-------
Section: 7(HD)
Revision: 0
Date: June 1975
Page 4 of 14
Relative humidity is almost universally determined in the
emission laboratory using the wet bulb-dry bulb hygrometer. Other
methods of determining humidity are available but attempts to correlate
the various methods have usually met with some unsolved discongruity.
Therefore, it is mandatory that the equipment used for humidity determin-
ation should be specified. Two basic types are presently used: the
fan-type hygrometer with either themocouples or thermometers and elec-
tronic or visual read out. The other is the sling-type psychrometer.
These two types are known to give equal readings.
A comparison of readings, on an audit basis, of these two
types could be used as a check. The sling psychrometer is the preferred
audit tool because of its portability.
Other recommended methods of reducing variability include a
controlled test lab environment, and continuous recording of humidity
during a test. Wicks and water supply should be inspected frequently
for contamination. Thermocouples and thermometers should have a calibra-
ted accuracy of ±0.5 F or better.
7.1.3 Barometric Pressure
The temperature-compensated aneroid barometers, calibrated
against standard laboratory mercury barometer are frequently used in
the measurement system. In laboratories with only a single test cell,
a mercury barometer is often used. The two primary sources of error
for barometer readings are error in calibrating the aneroid barometer,
and errors in the reading of a mercury barometer. Calibration errors
are generally controlled through independent checks. Errors in reading
the barometer can be reduced by recording the pressure before and after
the test. Comparison of the range of the readings could then be done
by data validation or computer, utilizing one of the control chart
techniques described in Section 6. In addition, comparison to the
reading of the previous test on the same day would provide an additional
check.
7.1.4 Fuel Measurement and Control
The fuel used in testing the engine is often overlooked as a
potential source of test error. Test laboratories usually have a
choice of three fuels popularly known as 91 Octane, Idolene 30 and
Idolene Clear (HO). Although the specifications for these fuels are
regulated by the EPA, this does not assure that the fuel obtained from
a supplier meet these specifications. The results of using leaded fuel
in a catalyst-equipped engine have been well publicized. Foolproof
controls must be implemented to preclude the use of the wrong fuel.
36
-------
Section: 7(HD)
Revision: 0
Date: June 1975
Page 5 of 14
Other characteristics known to have an effect on emissions
are the Reid Vapor Pressure (RVP), octane rating and hydrocarbon composi-
tion. The RVP can be changed through improper storing, overheating of
the fuel, age, and improper handling. The use of "weathered" fuel can
cause starting difficulties and, therefore, fresh fuel should always be
used for emission test.
In view of these potential sources of test error, fuel delivered
to a laboratory should be tested upon receipt for conformance to specifi-
cations and should not be released for use if the results of the test
differ from the specifications. Storage .drums should be clearly marked
and color-coded. Care must be taken to contain each type of fuel in
separate storage tanks, with thorough drainage of a tank prior to
filling with another type of fuel.
Using fuel of the wrong octane may cause "ping" or "knock" in
some engines possibly resulting in certification test failure. Hydro-
carbon composition, in part, determines fuel octane and the running
characteristics of the vehicle. In addition, the response of the FID
can be affected by different ratios of paraffins, aromatics and olefins.
Therefore, fuel analysis and correct fuel handling are important in
controlling test variability. As catalysts are added to control emissions
from gasoline engines, the lead content of the fuel becomes of prime
concern. Recently the National Bureau of Standards issued Standard
Reference Materials for the measurement of lead in fuel (SRMs 1636,
1637, and 1638), which provide a valuable tool for calibrating and
checking the lead analyses method used in fuel control systems.
The accuracy of the measurement of the mass of fuel used for
each mode will have a direct effect on the data. For example, a measure-
ment error of 2 percent will cause a corresponding error of 2 percent
in the determination of mass emissions in grams per hour.
Methods of fuel measurement may vary from one laboratory to
another as no single method has been specified. The method used should
be subjected to periodic calibrations as recommended by the manufacturer.
Fuel flow meters calibrated for a specific temperature and specific
gravity require corrections to be made to readings when fuel temperatures
and/or the specific gravity are significantly different from the calibra-
tion parameters.
Because of volatility and explosive nature of gasoline, "dead
weight" type measurement systems are difficult to use. Turbine and
displacement type meters are generally used, but care should be taken
to avoid vaporization of fuel in the system, which results in erroneous
readings and engine malfunction.
37
-------
Section: 7(HD)
Revision: 0
Date: June 1975
Page 6 of 14
7.2 VARIABLES ASSOCIATED WITH THE MEASUREMENT OF GASEOUS EMISSIONS
7.2.1 Variables Associated with the Analytical System
Exhaust emission concentrations are determined using an
analytical system calibrated with working gas mixtures which have a
specified accuracy of ±2 percent. Usually instrument curves are construc-
ted with calibration gas mixtures having accuracies of ±1 percent or
better. Gravimetric standards prepared and used by the EPA have a
reported accuracy of ±0.5 percent or better. In addition, reference
standards are available from the NBS (SRMs 1665-1669, 1673-1675, 1677-
1681, and 1683-1687). Instrument precision and reproducibility are
specified by the Federal regulations and, through experience, have been
found to conform to these specifications when properly maintained.
Successive analyses of the same sample give a precision of ±0.5 percent
of the full scale concentration (Reference 7-4).
The primary sources of variability in the analytical system
are:
o Accuracy of the calibration gases
o Instrument precision
o Accuracy of working or span gases
o Calibration curve construction
o Condition of the sampling system
o Full scale concentration
o Zero gas purity
o Instrument drift (electronic)
o Operator
The variables are controlled through a system of receiving
inspection, performance and audit checks, etc., described in an earlier
section. Detailed procedures for these appear in Volume II, the Test
Procedure Manual. In determining the effect of error in concentration
measurement, a coefficient of variation of 1 percent of the full scale
is usually encountered in a repetitive determination of the same sample
(Reference 7-3). However, variation between analytical systems has
been experienced as high as ±3 percent for the same sample. Correlation
values in excess of this are considered to be undesirable and suggest a
need for corrective action. Corrective action usually involves a
system leak check, reanalysis of the working gas and contruction of a
new instrument curve, followed by a systematic check of the sources
previously mentioned.
An error in the measurement of an exhaust gas component would
obviously have a corresponding direct effect on the mass emission
values. Measuring the concentration on the lowest convenient range
38
-------
Section: 7(HD)
Revision: 0
Date: June 1975
Page 7 of 14
improves the accuracy of the data. Instrumentation with a capability of
multiple range selection is available and used by many laboratories,
consequently they are able to select lower ranges than those specified
in the Federal Register.
Other sources which need further control are the instrument
zero drift, which should be checked periodically and contaminants in
the zero gases. Nitrogen and air zero gases should be rigorously
analyzed by the receiving laboratory in preference to the present
practice of accepting batch analysis from the supplier. The reduction
of contaminants in laboratory ambient air concentrations should also be
considered. Inspection of the heating system for leaks and proper
ventilation will help in achieving more desirable ambient conditions.
Because of the variety of available certified accuracies for
calibration gases, a decision must be made based on cost versus desired
reliability when obtaining the laboratory standards and "working gases."
Naturally, as the certified accuracy of the blend is improved, the cost
of the gas increases exponentially. In all cases, however, traceability
to the EPA primary standards, either through correlation programs or by
direct analysis by EPA, is desirable.
Calibration curves may be checked weekly by using the primary
set of gases. However, this is a rather lengthy process involving the
use of a large number of cylinders. When correct instrument maintenance
procedures are complied with, the instrument curve shape generally
remains stable for long periods of time. Consequently, a simpler
process would be to check the mid-point of the curve using a primary
standard and the span for the most frequently used ranges. These
results (the mid-point deflections) are plotted on control charts
(Section 6) for each instrument. When an "out of control" situation is
detected, the complete calibration curve should be repeated and the
span gas reanalyzed.
Nitrogen oxides (NO ) are presently being detected utilizing
the NDIR instrument which is specific for NO. Nitrogen dioxide (N0_)
is not detected by this method. NO in the presence of O_ will oxidize
to NO . Consequently, it is important when using this method to minimize
the sample transport time by locating the analyzer reasonably close to
the engine and using high flow rates. Future standards will most
likely be based on the use of a chemiluminescence analyzer using a
converter similar to those presently used in light duty testing. This
would eliminate the NO - NO_ conversion problem and the interference
from other exhaust components associated with the NDIR method.
7.2.2 Variables Associated with Data Reduction
The determination of emissions from heavy duty gasoline
engines is performed by running the engine through a series of cycles
39
-------
Section: 7(HD)
Revision: 0
Date: June 1975
Page 8 of 14
A. BSHC = V(HC x WF)
t-J mass
(Measured BHP x WF)
B. BSCO = (CO x WF)
mass
(Measured BHP x WF)
C. BSNO = VXNO x WF)
Z— < x mass
(Measured BHP x WF)
D. Average the composite BSHC, BSCO, BSNO
emissions of the first and second cycles.
E. Average the composite BSHC, BSCO, and BSNO
emissions of the third and fourth cycles.
F. Combine the results of (D) and (E) according
to the formula: 0.35 x composite of (D)
+ 0.65 x composite of (E).
G. Correct the BSNO value for the humidity
at test conditions by multiplying by
conversion factor "K" where:
K = 0.634 + 0.00654H - 0.0000222H3
H = Humidity at test conditions, grain
H O/lb. dry air.
It is apparent from these formulas that errors in the input
data would have a direct effect on the resultant mass emissions.
Weighting factors for each mode determine the magnitude of the effect
for each mode. Modes having the higher weighting factors would neces-
sarily have the greatest effect on the final numbers.
7.2.3 Computer
Computers, with their built-in checks and reliability, are
very useful in reducing test variability. The variety of computers used
in mobile source testing ranges from "desk top" to completely automated
systems. Although the computer is generally more reliable than manual
operations, it is not infallible and should be periodically checked for
reliability. One proven method of checking data reduction is the use
of a previously prepared, standard set of manually calculated data.
40
-------
Section: 7(HD)
Revision: 0
Date: June 1975
Page 9 of 14
This is fed into the computer and the output is compared with the calculated
values. The same set of data could be used in cell-to-cell or laboratory-
to-laboratory correlation studies.
7.3 THE ENGINE AS A SOURCE OF VARIABILITY
Heavy duty gasoline engines are themselves sources of varia-
bility on a test-to-test basis and during the determination of durability
emission data. Emissions are known to change significantly after a
break-in period and over the period of 1,000 hours of durability testing.
Consequently, deterioration factors are applied to the data submitted
to EPA for engine certification.
Engines are affected by ambient conditions and the applied
torque of the dynamometers. A reference or correlation engine should
be characterized by varying the various operating parameters within the
allowable operating ranges. A significant number of tests should be
run to characterize the emission profile.
Conditioning and soak time prior to a test are prime sources
of test variability. Engine soak time must be specified with ample
running time prior to soak to assure equilibrium conditions similar to
those encountered in running the engine for 125 hours intervals. The
Federal Test Procedure requires the preconditioning of an engine prior
to the start of the test. The definition of normal operating conditions
is unclear; many engine parameters not normally measured such as oil
and transmission temperatures may have a significant effect on emissions.
Better definition of this preconditioning procedure would reduce varia-
bility and improve correlation between laboratories.
All illustration of the effect of ambient conditions on
gasoline engine emissions are presented in Figure 7-1 and Table 7-1.
These illustrations were based on the light duty test cycle. However,
similar results could be expected with a heavy duty engine-dynamometer
test cycle. The engine, as a source of variability is difficult, if
not impossible, to control by the average emission test laboratory.
Engine variables are the responsibility of the manufacturer, but the
testing laboratory must assure that an engine installed in a test cell
is set to the correct engine operating specifications in order to
achieve a reliable determination of the emissions.
7.4 MEASUREMENT OF VARIABILITY IN THE EMISSION MEASUREMENT SYSTEMS
Variability of the measurement is defined as the -inability to
achieve identical test results from repeated tests on the same engine
without changes to hardware or engine adjustment specifications.
41
-------
Section: 7(HD)
Revision: 0
Date: June 1975
Page 10 of 14
-8
5
V)
UJ
s
3
s
S
UJ
»-
K
Z
UJ
m
<
10
UJ
K
QC
<
W
i
O
d
O
CO
>
0
Ollft/9) IHVJLS-CnOO SAO 'N30OU1IN dO S3QIXO
III
8
8 £
z
£
iS3i idvis-anoo SAO '
NOSUVO
i
In
oiiN/9) 1331 lavis-cnoo SAO 'NOSUVOOUQAH
42
H8
m
uj
5
O
UJ
LL
U.
UJ
UJ
O
UJ
cc
UJ
U.
UJ
cc
Lit
O
cc
D
O
CO
-------
Section: 7(HD)
Revision: 0
Date: June 1975
Page 11 of 14
Table 7-1. EFFECT OF BAROMETRIC PRESSURE
AND HUMIDITY ON EXHAUST EMISSIONS
PERCENT CHANGE
SOURCE HC CO NO CO RANGE OF STUDY
One inch Hg increase
barometric pressure
GM environmental
chamber data -10 -30 +5 +2.2 26-30 Hg
Ford data based on
multiple regression
analysis of three
vehicles -13.6 -21 +12.5 +7.7 28.7-29.51" Hg
50 grains increase in
absolute humidity
GM environmental
chamber data +10 +25 -1.5 30-100 grain/
Ib dry air
Data based on FTP-H tests
Source. Reference 7-9 p. 159
43
-------
Section: 7(HD)
Revision; 0
Date: June 1975
Page 12 of 14
Variability exists in test results to varying degrees dependent on the
type of variability, test-to-test, cell-to-cell within a laboratory, or
laboratory-to-laboratory.
A discussion of the importance of determining variability and
its effect on the automobile manufacturers has been presented by Ford
(Reference 7-4) and General Motors (Reference 7-5) in their applications
for suspension of the 1977 Federal Emission Standards. As the emission
requirements become more stringent, the level of variability significance
assumes more in affecting the ability to develop and certify emission
control systems. Variability factors are affected not only by the
vehicle, but also by the test-to-test variability. Determination of
the expected variability is important, therefore, to ascertain the
actual levels of exhaust emissions for certification of emission control
devices. These referenced reports consider both "in house" variability
and correlation factors which exist between the manufacturer's laboratory
and the EPA laboratory. Variability in emission measurement systems is
usually expressed as the coefficient of variation which is defined as
the standard deviation (s) divided by the mean of the results, expressed
as a percentage (CV = — (100 percent). Also, variability may be
defined for some confidence level; for example, to assess the variability
associated with a 90 percent confidence level, 1.96 is used as multiplier
in a similar calculation. In other words, as the confidence level is
increased, the confidence interval becomes wider. Therefore, in the
case of a certification engine, the higher the confidence level selected,
the more efficient the emission control system must be in order to
obtain the emission values required to be statistically confident that
all engines will meet the Federal Emissions Standards.
The Coordinating Research Council has carried out a four-
phase cooperative program to evaluate techniques in measuring gaseous
emissions in diesel exhaust (Reference 7-12, 7-13).
Each phase involved the measurement of diesel exhaust from a
single engine either by circulating the engine among the laboratories
or by all participants measuring emissions on the same engine at the
same time. Further details of this study are reported in the Phase II
Quality Assurance report. To our knowledge, no such data has been
published for gasoline engine emission testing laboratories. An important
function of quality planning involves studies of the relative contribution
of the sources of variability, and interlaboratory correlation programs.
Several statistical methods have been designed for performing these
studies with a minimum of data. In addition to the statistical methods
discussed in Section 6, two other references should be consulted;
(1) a recently published "Handbook for Air Pollution Measurement"
(Reference 7-10), and ) a document prepared by W.J. Youden for the
Association of Analytical Chemists (Reference 7-11).
44
-------
Section: 7(HD)
Revision: 0
Date: June 1975
Page 13 of 14
7.5 QUALITY ASSURANCE AND TEST VARIABILITY
Statistical methods for use in controlling test variability
have been described extensively in Section 6. Quality assurance has
the responsibility for control of test-to-test variability and improving
data reliability. Many studies have been completed on methods for
reducing test variability. However, further reduction of test variability
is impractical in many cases; consequently, quality assurance should
advocate the use of procedures such as data validation, calibrations,
and maintenance, and assure that these procedures are being complied
with. Table 7-2 is a summary of the test variables and the methods
used for their control.
Many of the precautions and checks discussed in this section
are included in the Test Procedures (Volume II). Each test facility,
depending upon its experience and judgment, should carefully review
this section to determine if some or all of the additional precautions
and checks should be introduced as routine or periodic checks into
their operational test procedures.
45
-------
CN
i
r-
cu
rH
XI
(C
EH
4->
CO
0)
EH
M-l
0
^i
J^l
(0
H
P
CO
CO
'O
O
x;
4J
0)
s
TJ
G
fl
CO
cu
X!
fC
•H
j^
IB
>
O
M
4-1
G
0
U
^1
•H
CU
x:
EH
^_J
O
Pn
r0
0)
CO
D
CU
W
D
13
.SS
4-1 C
0) 0
S O
CO
cu
EH 0)
CU
-I rH
P3
<0
>
Section: 7 (HD)
H
•^
r
^>
^
V,
*>>
^.
f
•>»
f
^v
^,
f
~f
**^
r
•*+*
7
•**.
If
^
•s.
co
s
i^j
H
p^
^
EH
C/3
H
levi
)ate
-
-
-
x
X
X
X
CU
•H
Oi
W
>ion
14
X
X
X
X
X
X
}_l
dJ
.p
QJ
namom
Q
: 0
J
>f 1
X
X
X
X
X
X
X
^!
+J
•H
•O
•H
K
me
X
X
X
X
X
X
X
M
Q)
romet
HJ
.975
X
X
X
to
G
0
-H
4J
•H
•O
G
0
4->
G
CU
•H
i
X
X
X
X
X
X
X
X
X
X
X
M
0)
N
rH
(0
c
rt
X
X
X
X
X
c
0
•H
4-1
libra
id
o
X
X
X
X
X
to
o
B
0)
N
X
X
X
X
en
(0
o
G
•H
•a
o
2
X
X
X
lj
erato
&
X
X
X
X
X
^1
0)
4-1
1
0
U
X
X
X
X
X
X
X
14
CU
4-1
01
rH
0)
X
X
X
X
X
X
X
X
e
0)
CO
co
tr
G
•H
rH
ft
W
X
X
X
X
X
CO
^1
-------
Sections: 8(HD)
T, ' '
Revision:
Date: June 1975
Page 1 of 3
Section 8
QUALITY ASSURANCE SYSTEM SURVEY
REFER TO PHASE II REPORT
Section 9
REFERENCES
REFERENCES INCLUDED FOR SECTION 7 ONLY
OTHER REFERENCES AS LISTED IN PHASE II DOCUMENT
Appendices, A, B, and C
REFER TO PHASE II REPORT
-------
Section: 9(HD)
Revision: 0
Date: June 1975
Page 2 of 3
REFERENCES PHASE IV
7-1 Guidelines for Quality Assurance Programs for Mobile Sources
Emissions Measurements Systems - Phase I. EPA Report No.
650/4-75-024, Research Triangle Park, North Carolina, June
1975.
7-2 Consultant Report to the Committee on Motor Vehicle Emissions
on Emissions and Fuel Economy Test Methods and Procedures,
Washington D.C., September 1974, Section 4.4.
7-3 Paulsell, C.D. and Kruse, R.E., Test Variability of Emission
and Fuel Economy Measurements using the 1975 Federal Test
Procedure, Society of Automotive Engineers, Inc., New York.
Publication No. 741035.
7-4 Application for Suspension of 1977 Motor Vehicle Exhaust
Emission Standards. Ford Motor Company, Volume 1, Section III-
E, January 1975.
7-5 General Motor's Request for Suspension of 1977 Federal Emis-
sion Standards, Appendix 20, Volume III of III, January 10,
1975.
7-6 Klingengerg, H.; Fock, M.; Lies, K.H.; and Pazsitka, L. "A
Critical Study of the United States Exhaust Emission Certifi-
cation Test-Error Analysis for the Test Procedure," presented
at the 67th Annual Meeting of the Air Pollution Control Associa-
tion, Denver, Colorado, June 9-13, 1974.
7-7 Fock, M.; Lies, K.H.; and Pazsitka, L.; "Critical Study of
the United States Exhaust Emission Certification Test-Error
and Probability Analysis," presented at the Society of Auto-
motive Environs Meeting, Houston, Texas, June 3-5, 1975, Paper
No. 750678.
7-8 Diesel Egine Emission Measurement Procedure, Society of
Automotive Engineers, Inc., New York, SAE J1003, SAE Recom-
mended Practice, March 1973.
7-9 Report by the Committee on Motor Vehicle Emissions, Commission
on Sociotechnical Systems, National Research Council National
Academy of Sciences, Washington D.C., November, 1974.
53
-------
Section: 9(HD)
Revision: 0
Date: June 1975
Page 3 of 3
7-10 Quality Assurance Handbook for Air Pollution Measurement
Systems, Environmental Protection Agency Report (Preliminary
Draft) Appendix K, Research Traiangle Park, North Carolina.
7-11 Statistical Methods For Analytical Chemistry, W.J. Youden,
Published by the Association of Official Analytical Chemists,
Washington D.C. 20044.
7-12 Wagner, T.O.; Broering; L.C.; and Johnson, J.H.; CRC Evaluation
of Techniques for Measuring Hydrocarbons in Diesel Exhaust -
Phase IV, Society of Automotive Engineers, Inc.; Warrendale,
Pennsylvania. Publication No. 750203, February, 1975.
7-13 Perez, J.M.; Broering, L.C.; and Johnson, J.H.;, Cooperative
Evaluation of Techniques for Measuring Nitric Oxide and Carbon
Monoxide (Phase IV Tests), Society of Automotive Engineers,
Inc.; Warrendale, Pennsylvania. Publication No. 750204,
February, 1975.
54
------- |